My overall PPC strategy - I want to share/discuss it with you

You don't really need keyword tools all that much. Let's say you sell office chairs. How many different ways can you say that? Keyword tools won't likely help much, if at all. Google's tool will be enough if you are so inclined to check. It may help spot negatives you haven't thought about but those will quickly be apparent as you gather actual campaign data.

Your third step is therefore not really necessary. You know you'll get thousands of keywords, most are not appropriate so why put yourself through that? At least, I would not go through thousands, check the first dozen or so but don't dwell on this. You should have your main keywords in your first step.

Now, there are always exceptions of course. There are only a few possible keywords for most products. But in general, you'll have just one group for each. Even services will have just one or two. I manage a campaign for a hair transplant doctor and there are just two main ways to say this (in French). So basically, your steps 2 to 4 are just one with little or no need for any tool other than your brain.

Match types: no, I don't think you are wrong to act so widespread. I do this. Just use the broad match modifier using + in front of words, not the simpler broad match. This applies for both Adwords and Bing Adcenter.

Come up with as many ad ideas as possible. But only use two active ads at the same time.

Setting low bids. This should be relative, not a fixed amount. In my view, you want to reach the most people searching on your terms as possible. I think most people want that too. This means you want a reasonable position and that means at or near the top, positions one to three. This means you have to bid the going rate. This way, you maximize the number of clicks you get which is the whole point.

However, I do this only for great quality score, seven or more. As someone who does this all the time for a living, I usually get there on the first try. That may not be the case for you if you are starting out. In fact, chances are you will achieve only average QS or less which will make your costs higher. But you still need to get exposure and learn how your ads are fairing.

I still think you need to bid the typical amount for your keywords. So if the typical bid is $1, bid that or nearly that amount. I would say bid no less than 75% of that. Improve your campaign to improve positioning and keep costs in check or lower relative to competitors.

I like to have control and not some algorithm decide things for me.

Optimization is mostly about improving your ads to get more clicks (and it may help with conversions too). You need enough data before making a sound decision. This means enough clicks, forget impressions or length of time. You need at least 20 clicks per ad but more is better. It could take a day or it could take months.

Some times, you have an ad that is doing so poorly, you don't wait until you have enough clicks. You'll see it in other data such as the ad getting a lower positioning of a few positions. That tells you it's not as good. You may also see a higher CPC. When I see odd things like that, I stop that ad even if I don't have the required number of clicks.

I don't understand your reasoning behind your formula. First, why the multipliers? Why those specific ones? Do you have any data to back them?

Since most of QS is CTR, I don't see the point of including both, much less multiplying them. You are also applying it to keyword data. It should be on the ads since they are driving the clicks. Any analysis should be on the ads. I have a background in databases and don't use spreadsheets, using more sophisticated approach. This is out of reach for most so I understand using spreadsheets instead but they are inadequate for all but simple analysis.

You are forgetting one thing: optimizing your landing page to increase your conversion rate. Getting quality clicks is good and the first step. You need to close the sale. Look at both sides of the coin.

>> With a tool you can find keywords that you can't imagine

You need a tool to come up with different ways for "office chair"?

There can be so many different misspells, I don't think it's worth to try to find them all. Besides, they are rare, relatively speaking. You'll rarely get misspells that have high search volumes, I can't think of any off hand. Also, Google takes care of them very often and the benefits of trying to find them is so minimal it's not worth it. At least, not to me when I handle many accounts.

You say you work with services. It's still fairly simple in my view what the keywords should be in most cases.

If the service being offered is hair transplant operation, that will be your primary keyword. You mention hair loss which could be a secondary keyword in this case and may be worth a try but in my experience, will have much poorer ROI. That is too generic anyway. A better one is "hair loss replacement". Wig is not a keyword I would use in this case. That is not the product or service being offered. You may say I'm missing out on potential conversions and that may be true. But those will be more expensive as those looking for a wig are much less likely to go for a more expensive option, a surgery no less.

>> why not starting with a very low bid and wait Google for tell us the minimum offer for the first page?

You should have a fairly good idea of the going rate for a keyword. The keyword tool will practically tell you what that is.

Why wait for the system to tell you? This is going to be based on your QS anyway. If you have a poor one, your first page minimum bid will be higher. Improve QS and it will be lower. I don't need to bid low to find that out.

>> 200 impressions

Look at it this way. You may get a 1% CTR. That's just two clicks. You'll make a decision based on two clicks? One ad may have an extra click. You can't compare a 2/200 with a 3/200. You just don't have enough confidence that the second is truly your best, the next 200 impressions may be just the reverse.

>> I've choosen the multiplier following my intuition

There should be little intuition here. You need science. Empirical data. I don't think your formula is of any value.

>> The fact is that some time I don't see a strong correlation between the two parameters.

Because you are looking at absolute CTR, not relative CTR. QS is a relative calculation. Relative that is compared to what your competitors have accomplished. That will be different for each keyword. And since you don't know this historical data, you cannot say that a 2% CTR is equal to a certain QS. It doesn't work that way. Besides, QS also takes into account position. So don't listen to those that say that a certain CTR is good or bad. You don't know the position or what competitors are doing. QS is telling you how well you are doing as it is a value that takes all this into account.

You also have CPC in your formula. That is irrelevant.

What I do in effect to give me a better understanding of which is a better ad is my own QS calculation based on CTR and position. I compare my own ads since that's the only data I have and it's worked well. For campaigns tracking conversions, I incorporate that into my calculations and take into account a few other things as well in order to compare apples to apples such as any change in the campaign.

You said:

3) CREATE A LIST WITH *ALL* THE WORDS SUGGESTED BY GOOGLE
I put the words found in step 2 in the Keyword Tool Planner. One by one. For each of them I fetch ALL the suggested words in the first 3-5 pages, using the "Copy to Clipboard". Then I get a txt with thousands of words, some of which are duplicate or bad - does not matter for the moment.

Why don't you check the ad groups in keyword planner instead of keywords. In the ad group you will get similar keywords. Another thing, why do you use "Copy to Clipboard" method instead of extracting those keywords ideas in .CSV format? It will be helpful to filter out unnecessary keywords (You can define unnecessary keywords as red in CSV and save that CSV to .XLSX after the modification)

5) APPLY 3 MATCH TYPE
Each word is changed via 3 different match type: modified broad / phrase match / exact match. So the number of words in groups triples.

So here is what you did right?

+Keyword 1st
"Keyword 1st"
[Keyword 1st]

It's not a good idea to add three match types for keyword because these keywords (match types) will compete with each other for higher quality score which can increase CPC for that keyword.

There is no problem having all three match types together in a group. They don't compete with each other. The system will use the keyword with the best option based on your match types and the search query. This regardless of whether your keywords are in one group or three.

The system looks for any exact phrase matches to the search query. It will use it first. If you don't have any, it then checks for phrase match. If none, it then goes to broad matches, the modified version first then finally the plain old broad. There is no competition between keywords and this part is not affected by QS. It is simply the first step in determining if there are any hits between the search query and advertisers' keywords.

Please exscuse me for the late response. I'll read carefully all your answers and I'll come back here in a couple of day.